If you’ve been in supply chain for any length time, you might be wondering what caused demand forecasting to develop so many different ‘personalities’ over the years, such as demand forecasting, demand planning, demand sensing, and demand shaping. What’s going on?
It’s easy to ignore all these as buzzwords, but that would be overlooking the genuine progress that has taken place over the past several decades. The days of simply using statistical methods to forecast future demand using past sales history are fading. Computing power and storage capacity have grown exponentially, while the cost of both have plummeted. More and better data has turned demand analytics into mainstream reality.
As a result, a whole new language has developed around the demand forecasting process that can be confusing. Let’s take a few minutes to decode it.
The art and science of forecasting customer demand to optimize supply decisions.
Demand forecasting describes the decades-old science of predicting demand. Techniques range from informal methods like educated guesses to quantitative methods that analyze historical sales data. It has matured as new algorithms and techniques were introduced that took into account lost sales, seasonality, lumpy demand, promotions, etc. It often employs statistical metrics like MAPE (mean average percentage error), which has hit a wall in recent years due to increased demand volatility and this approach's mostly backward-facing nature.
A multi-step operational supply chain planning process used to create reliable forecasts.
The approach begins with a statistical forecast. Data sources for the forecast can include planned sales orders, customer contracts and intercompany standing orders. The final forecast is shared with key stakeholders, such as sales and suppliers, for their input.
Demand planning improves on demand forecasting in two key ways. First, it is more collaborative. It typically involves meeting with multiple stakeholders and building consensus forecasts. Demand Planning often supports sales and operations planning (S&OP) initiatives. Second, it can be a continuously improving process, where planning outcomes are regularly assessed and used to fine-tune the next cycle using tools such as Forecast Value Add (FVA).
A demand forecasting method that leverages mathematical techniques and near real-time data
to improve the forecast based on supply chains’ recent realities.
Demand sensing – which can also referred to as Demand Analytics - is a structured, automated process by which downstream and third-party demand signals are captured and analyzed to inform the latest demand forecast. Demand signals can include downstream demand such as “sell out” or POS data and downstream inventory levels. It can include outside demand signals such as weather, new housing starts and other macroeconomic indicators. And it can also include ‘big data’ (often non-relational or unstructured) such as clickstream data. Demand sensing software must reconcile the signals with the forecast.
Because demand sensing builds on demand planning by leveraging more data types, it often relies on machine learning to handle the volume and variety of data involved and differentiate meaningful signals from noise. Like demand planning, it is usually a continuously improving process.
The influencing of demand to match planned supply.
To shape demand, companies have traditionally used tactics such as price incentives, cost modifications, promotions and product substitutions to entice customers to purchase specific items when demand and supply plans are not aligned. It’s ideal if the demand elasticity for levers, such as price, can be calculated allowing companies to fine-tune outcomes to minimize waste and maximize profits.
Teams using advanced analytics technology usually reduce the need for demand shaping as a corrective measure. However, as long as random events occur to create unexpected changes to supply and demand, demand shaping will remain an important cross-functional planning process.
The barriers to acquiring the advanced analytics technology are at an all-time low, allowing businesses to look forward, anticipate and respond to demand changes in real time. If you are still planning your demand by looking in the rear-view mirror, there’s never been a better time to evolve.
Click below for a copy of our popular "Evolution of Forecasting" infographic: